Is a PON necessary for tumor-normal matched Mutect2?

Question

[sarahgzb]

I'm a bit confused on whether or not i should include GATK's public PON (either 1000g_pon.hg38.vcf.gz since I aligned with hg38), make my own from my normal samples, or just leave it and not include a PON. I am planning on including --germline-resource af-only-gnomad.vcf.gz

[Zhenyu Zhang]

of coz you still need it even you have gnomad included.

Aligned and variant calling algorithms are not perfect, and the PoN is needed to capture all potential artifacts.

as to the options, your own PoN is always the best option, followed by public PoN.

[LChart]

OP has matched normals. PoN is a solution for when paired normal is not available.

[Zhenyu Zhang]

This is simply not true. PoN is needed to capture artifacts even when paired normal and germline database are already used

Answer 1

[LChart]

If you have matched tumor/normal samples, you do not need a PON. That said, I don't think adding one would hurt (except for runtime).